Rooting out the bad apples

Avocados at Granot, Israel’s largest avocado grower, are basically supermodels. Each one is meticulously photographed from dozens of angles, searching for small blemishes. This allows the cooperative to sort the avocados by quality and command higher prices.

But even supermodels hounded by the paparazzi can fail to deliver. Everyone knows the disappointment that comes with opening an avocado that looks and feels perfect, only to discover it is mottled with black inside instead of the bright green creaminess.

“When you only see the fruit from the outside, we don’t know when it will be ripe, or the quality of the interior of the fruit, which makes it really hard to predict how it will behave during the shipping and how it will be at the customer’s house,” explains Yakov Armon, R&D manager of Granot Fresh.

Granot is Israel’s largest avocado grower and exporter, an agricultural cooperative of 42 kibbutzim and moshavim. In 2021, the company grew 30,000 tons of avocados.

“Even though we’re already using so much technology and automation, we really only know what the outside of the fruit is like and its weight,” Armon says.

Avocados can be finicky in their ripening process, leading to huge waste. In an effort to address the problem, Granot turned to Neolithics, an Israeli startup that uses hyperspectral optics and artificial intelligence to reveal the secrets of the inner life of fruits and vegetables.

“When we use Neolithics, we’re able to sort the avocados according to when they might ripen, so whenever people open an avocado it’s perfect,” said Armon.

Using Neolithics’ scanning technology, Granot can now predict with up to 90 percent accuracy when they believe an avocado will be ripe, enabling the company to supply different stores with different internal supply chains and varying times until the avocados are placed on the shelves.

“We want to bring the avocados to the retailer at the exact ripeness they want,” said Armon.

“Fresh produce is the most difficult thing for grocery store management because of the variants and the changes along the supply chain,” explains Amir Adamov, the CEO of Neolithics. “At least 20 percent of fresh produce is lost in the supply chain for many reasons.”

Neolithics is also working closely with Israel’s biggest retailer which uses the technology to help determine the best distribution schedule for dozens of fruits and vegetables.

The retailer was able to slash its monthly quality control budget by more than 80 percent using Neolithic’s scanning technology at its warehouse.

Neolithics uses hyperspectral cameras to inspect produce from 360 degrees, looking for signs ranging from those visible to the naked eye, like bruising, to internal chemical properties like sugar content.

Neolythics’ hyperspectral cameras reveal the inner life of nectarines and other fruit and vegetables, helping stores to check for freshness (Neolithics)

The AI compares each scan to a massive trove of images, determining which produce meets the standards for firmness, predicted shelf life, color, or other significant properties. A store will be able to judge which produce is closer to spoiling and should be first, or at a discount, and which will stay fresh for longer.

It’s not just about profits. About one-third of all food in the US is wasted, costing billions of dollars and creating a carbon footprint greater than the airline industry. Globally, wasted food is estimated to account for about 8 percent of greenhouse gas emissions.

“There’s lots of inequality in the world about food distribution and consumption: some get a lot and generate a lot of waste, and some get nothing,” said Adamov. “We thought, there must be a way, with technology, to improve this process.”

“We care about this issue. We’re foodies. We hate to see good stuff go to waste,” Adamov says.

Quality control in the produce supply chain currently depends largely on workers who inspect fruit to identify spoilage or manually chose a few items from each shipment for lab testing to ensure quality. Tests take a long time and only sample a fraction of the produce, while human inspection is inefficient for large amounts.

Grocery stores often sell hundreds of different lines, making it difficult to have a one-size-fits all approach to testing for freshness or shelf life.

Quality control currently costs about 6 euros per ton of produce: 1 euro for inspection, 3 euros for lab tests, and 2 euros for produce automatically lost to spoilage.

The Neolithics platform reduces the cost to just 1.5 euros per ton for the basic feature of determining freshness. There are also options for more advanced features, including predicting shelf life or identifying the presence of pesticides.

Technology is developing so fast that even three years ago the type of data analysis performed by Neolithics would have been prohibitively expensive, Adamov says. Computers can now handle immense quantities of data, allowing the technology to scan every single piece of produce passing under the hyperspectral cameras in real time.

“Now what we can do is take all of the information into account and decide: this apple is good or bad, it will last three days on the shelf, it has such and such sugar content, starch, and acidity,” explains Adamov, who has been interested in the combination of power of high-powered optics and AI since his days as an officer in the Israel Air Force.

“It teaches you to have fast and accurate situational analysis, because you have to make quick decisions with fast execution,” he says, comparing it to running a startup business.

The company is now raising its first venture capital round from investors and via the Jerusalem-based OurCrowd investment platform. The funds will be used to initiate direct sales in the US and Europe, and continue research and development.

“We’ve proven ourselves in Israel and now we want to explode,” Adamov says.

For more information about investing in Neolithics via the OurCrowd platform, click HERE.